srtm data analysis

3385
6
07-21-2014 08:08 PM
Divya_UdayanJ
New Contributor

I need help to analyse the SRTM DATA (1025 X 1025) of a particular terrain which has gaps or voids.

Could anybody suggest a solution to find the percentage of gaps or voids in the SRTM data? I would like to know if there is a tool for such data analysis?

It is also possible for me to share the data if needed.

The file format is (.geotiff)

0 Kudos
6 Replies
DanPatterson_Retired
MVP Emeritus

There is no specific tool, however, the spatial analyst extension has a variety of tools that can be used.  There are questions that need to be answered.  These gaps etc, are they truly nodata values or are they areas of cloud reflectance?  Have you examined the statistics for the dataset?

0 Kudos
Divya_UdayanJ
New Contributor

Thankyou for the reply.

I have done an analysis in Global mapper by creating a contour map of the dataset and analyzing it.

I am hereby attaching the video of the analysis. I am not sure, whether we can infer anything from the elevation values.

I doubt whether '0' value resembles nodata?

0 Kudos
DanPatterson_Retired
MVP Emeritus

Don't see an attachment...anyway any elevations that seem spurious are most likely cloud reflectance, for example, if the elevations in your area are all less than 1000m and you are getting values of 2000+m then these values are probably in error...the higher the value, the more likely they are due to reflectance issues

0 Kudos
DanPatterson_Retired
MVP Emeritus

file still can't be downloaded...is this thread not finished?

0 Kudos
curtvprice
MVP Esteemed Contributor

SRTM data has gaps mostly from terrain effects (radar shadows) - not [usually anyway] the atmosphere.

If you can identify the cells that are "gaps" by their value (poke around with the identify tool) then you can classify your tiff into two classes like this using the Raster Calculator tool:

Con("srtm_raster" > 1000, 0, 1)

You can then convert the raster to polygons and from there you can calculate areas. You may need to convert multipart polygons to single part (there's a tool for that) to get areas for each individual blob.

I'm assuming your tiff doesn't have NODATA cells. If the gaps are not crazy values but actually classified as NODATA, you would then do it this way:

Con(IsNull("srtm_raster"), 0, 1)

0 Kudos
DanPatterson_Retired
MVP Emeritus

you can also use the con reclass approach to get rid of extraneous values, then  Regiongroup  the areas and use Nibble to bleed values in from the edge of the holes  This is documented in the help files